Hi to u all, Last time i was reading about FM and its inventor,major E.Armstrong in an American History magazine.He introduced FM to the USA and thereafter to the whole planet.His superhet system is still used to this day in almost all communication equipment worldwide.FM was better than AM as far as noise performance was concerned althoughit was a more complex system .FM stereo which was developped in the late fiftees brought a new dimension to audio and shrewed maufacturers proposed equipment to get the most out of it.
Because of channel spacing,the federal authorities had limited the sound of FM stereo broadcasting to 15Khz instead of the full20Khz bandwidth.The main reason was to preventovercrossing to adjacent channels and also because the healthy human cannot perceive sound beyond 10Khz.yet some experts are saying that this 15Khz limit is detrimental to the sound and harmonics which contribute to the beauty of the audio are lost in the process.
1)Is that affirmation true?
2)by multiplexing the audio to the stereo decoder,the sound is further deteriorated .Is that also true?
3)By building two mono transmitters to do the job,can one obtain a better sound? 4)Lastly.Are there new technologies that can dispense the use of multiplexing the sound? Awaiting for your comments and observations,
Regards,
Revenant.
Because of channel spacing,the federal authorities had limited the sound of FM stereo broadcasting to 15Khz instead of the full20Khz bandwidth.The main reason was to preventovercrossing to adjacent channels and also because the healthy human cannot perceive sound beyond 10Khz.yet some experts are saying that this 15Khz limit is detrimental to the sound and harmonics which contribute to the beauty of the audio are lost in the process.
1)Is that affirmation true?
2)by multiplexing the audio to the stereo decoder,the sound is further deteriorated .Is that also true?
3)By building two mono transmitters to do the job,can one obtain a better sound? 4)Lastly.Are there new technologies that can dispense the use of multiplexing the sound? Awaiting for your comments and observations,
Regards,
Revenant.
There is a 19kHz pilot tone, used in decoding the stereo information,
that is notched out by a filter, which limits the useful audio bandwidth.
The stereo information is encoded in AM (not FM), on the 38kHz DSBSC
subcarrier, so there is more distortion and worse noise for stereo FM.
Last edited:
People argue about whether we need more than 20kHz, yet few people can hear much beyond 16kHz. If we were starting from scratch now, FM might have a slightly wider bandwidth but 16-17kHz is what we have and that is enough for good fidelity.
FM stereo is a compromise, but a very successful compromise given a good tuner and a suitable antenna. Not all tuners and antennas are good enough.
Using two mono transmitters would almost certainly give worse sound than FM stereo because you could not guarantee identical paths so channel balance would vary.
Multiplexing (for the reason I just said) is the right way to do it. How you do the multiplexing could change but there is no good reason to do so.
In theory a good digital system could be better than FM. In reality the broadcasters are mostly not interested in good sound so digital systems typically give worse sound than FM.
FM stereo is a compromise, but a very successful compromise given a good tuner and a suitable antenna. Not all tuners and antennas are good enough.
Using two mono transmitters would almost certainly give worse sound than FM stereo because you could not guarantee identical paths so channel balance would vary.
Multiplexing (for the reason I just said) is the right way to do it. How you do the multiplexing could change but there is no good reason to do so.
In theory a good digital system could be better than FM. In reality the broadcasters are mostly not interested in good sound so digital systems typically give worse sound than FM.
Actually the Dutch radio pioneer Hans Henricus Schotanus a Steringa Idzerda already used a form of FM modulation from his very first broadcast onwards, which was on 6 November 1919 at 8 PM Dutch time. It was a form of narrowband FM, though, that didn't have any of the advantages of Armstrong's system. Idzerda's FM system allowed a very simple transmitter construction.
Superheterodyne reception was invented almost simultaneously by the Frenchman Lucien Lévy, leading to a series of legal procedures that were eventually mostly won by Lucien Lévy. I must say that the description in Armstrong's patent is a lot clearer than the one in Lévy's patent.
I've read some very contradictory things about that. In a 1950's article from the BBC, they conclude after a controlled listening test that only a few of their most experienced people could hear any difference at all between audio band limited to 12 kHz and not deliberately band limited. On the other hand, a 1990's Japanese article indicates that signals above 26 kHz affect the brain waves of Japanese gamelan players listening to a gamelan recording.
I think the consensus is that 15 kHz is a bit low for the bandwidth of an audio chain, even though most people won't notice the difference.
The main issue with FM stereo is the increased noise when you have weak reception. The reason for this is a phenomenon known as triangle noise: an FM demodulator converts white noise in the channel into noise that increases with 20 dB/decade at the demodulator output.
For FM mono, at the transmitter side, the audio signal is passed through a pre-emphasis filter that amplifies the higher audio frequencies more than the low frequencies. In the USA, the gain is essentially flat up to 2122 Hz and then increases by 20 dB/decade. In western Europe the corner frequency is 3183 Hz. You can get away with this because the high frequencies are usually much weaker than the low frequencies in audio signals.
After demodulation in the receiver, the FM mono signal is passed through a de-emphasis filter that does the opposite: attenuate the high frequencies from 2122 Hz or 3183 Hz onwards. As a result, it also much attenuates the triangle noise.
For FM stereo, left and right are first passed through pre-emphasis filters (and sharp low-pass filters at 15 kHz), then the difference between left and right is double-sideband modulated on a 38 kHz subcarrier. The average of left and right (that is, the mono signal) and a relatively weak 19 kHz pilot tone are added to the DSB signal, producing the so-called multiplex signal. The multiplex signal is then FM modulated onto an FM carrier.
This means that the whole left minus right signal is centred around 38 kHz before FM modulation and right after FM demodulation. As a result, it suffers much more from triangle noise than an FM mono signal. The pre-emphasis/de-emphasis is also not nearly as effective as it is for FM mono.
I think that if FM mono had not been invented first, every FM stereo transmitter would consist of two separate FM transmitters, one for left and one for right, or of some construction that produces the same output signal. Both transmitters would have to use the same aerial (antenna) for the reasons DF96 mentioned, but that could be arranged. It would give a far lower noise during poor reception than the present system. However, FM mono had been invented first and the new stereo system had to be compatible with the equipment and channel allocations that were already in use.
Last edited:
!I'm pretty sure that two carrier frequencies close to each other will work well enough in practice, because that's what was used for Dutch stereo television until everything went digital. The (L+R)/2 was frequency modulated on a carrier at 5.5 MHz offset from the picture carrier and the (L-R)/2 on a carrier at 5.74 MHz offset from the picture carrier. They had to use sums and differences for compatibility with mono televisions that only received the signal at 5.5 MHz offset, I don't know whether they also did it to minimise the subjective impact of multipath interference.
In any case, the main thing is that they avoided the 22 dB(A) noise penalty of FM stereo by not putting anything on an ultrasonic subcarrier and putting that into an FM modulator.
Digital. MP3 and similar. You can put several programs on one carrier with "little" loss. "Digital FM". It has been done in Europe. Not everybody is a fan. Perhaps because with new digital flexibility the broadcasters get greedy and put more than a few signals on one carrier and let the quality slide. (Also takes sum+difference and often "simplifies" the difference to the detriment of stereo.)
5KC 7KC 10KC 12KC 15KC and 20kHz have all been "good enough" to different generations. A lot depends on signal to noise. When we could barely get 40dB S/N, anything over 7KC was hiss, and better to cut it. Experienced listeners preferred this to wide-range with hiss. In AM the S/N is all about transmitter power to local lightning and cross-mod power, and even 40dB in local reception requires HIGH-power transmitters. The biggest would throw signal halfway around the world, causing international friction. FM separates S/N from raw power by taking a wider bandwidth. The 15KC in 150KC selection (not practical in the 500-1600KC band) gives a reasonable balance of S/N and audio bandwidth and spectrum space. Going to 20KC audioband implies 3/4 as many channels in the same spectrum space. While spectrum use may be light in Mauritius, in most of US/EU it is extremely tight, signals shoulder-to-shoulder interfering with each other.
Of course. I don't know whether that would have been considered a big deal when FM broadcasting started, but assuming that the bandwidth per FM signal would have to be reduced to compensate for that, it is still better than the system we have now.
According to Carson's rule, you need about 180 kHz for FM mono and 256 kHz for FM stereo with the present standards.
Now suppose you have a bandwidth of 256 kHz at your disposal to distribute between two FM signals, one for left and one for right or one for (L + R)/2 and one for (L - R)/2, depending on what sounds best under multipath conditions. Each signal could then use 128 kHz. In reality you need some distance for filtering, so make that 100 kHz, leaving 56 kHz of distance. (You don't need much suppression because the signals are about equally strong.)
With 100 kHz bandwidth and 15 kHz maximum audio frequency, each FM signal can have 35 kHz peak deviation according to Carson's rule. Compared to the 68.25 kHz we have now for audio (the rest is used for the pilot tone), that's a loss in SNR of about 5.8 dB. (With separate left and right channels, the loss for mono reception would only be 2.8 dB due to the averaging of left and right.)
With the present multiplex system, the noise floor comes up by 22 dB(A) assuming 50 us de-emphasis when switching from mono to stereo. The difference only decreases to around 5 dB when the reception is very good indeed, so LO phase noise and demodulator noise dominate over channel noise.
One issue that might have been a problem with such a system is the required tuning accuracy. Then again, it is not that much worse than for the FM system we have now.
According to Carson's rule, you need about 180 kHz for FM mono and 256 kHz for FM stereo with the present standards.
Now suppose you have a bandwidth of 256 kHz at your disposal to distribute between two FM signals, one for left and one for right or one for (L + R)/2 and one for (L - R)/2, depending on what sounds best under multipath conditions. Each signal could then use 128 kHz. In reality you need some distance for filtering, so make that 100 kHz, leaving 56 kHz of distance. (You don't need much suppression because the signals are about equally strong.)
With 100 kHz bandwidth and 15 kHz maximum audio frequency, each FM signal can have 35 kHz peak deviation according to Carson's rule. Compared to the 68.25 kHz we have now for audio (the rest is used for the pilot tone), that's a loss in SNR of about 5.8 dB. (With separate left and right channels, the loss for mono reception would only be 2.8 dB due to the averaging of left and right.)
With the present multiplex system, the noise floor comes up by 22 dB(A) assuming 50 us de-emphasis when switching from mono to stereo. The difference only decreases to around 5 dB when the reception is very good indeed, so LO phase noise and demodulator noise dominate over channel noise.
One issue that might have been a problem with such a system is the required tuning accuracy. Then again, it is not that much worse than for the FM system we have now.
Last edited:
From a business perspective, it's hard to imagine something more unappealing than a system that requires twice the transmitters, twice the spare parts and standbys, twice the maintenance, and twice the electricity as the guy running a mono setup. Worse, it's incompatible - you still need the mono transmitter, so the costs are 3x. Will the stations customers pay 3x for ads? Of course not.
In the receiver, twice the gangs in the tuning capacitor, twice the local / conversion oscillators not just in close physical proximity, but close in frequency. What could possibly go wrong?
Will Joe and Jolene six pack appreciate what ever slightly improved fidelity such a system might or might not provide? I'm going to go with no.
On the other hand, the existing system is mono compatible, the increased costs are slight and a mostly one time expense - you don't even have to worry about the stability of the BFO in the receiver, because you get it from the transmitter and double it.
The system in place seems more brilliant than flawed.
In the receiver, twice the gangs in the tuning capacitor, twice the local / conversion oscillators not just in close physical proximity, but close in frequency. What could possibly go wrong?
Will Joe and Jolene six pack appreciate what ever slightly improved fidelity such a system might or might not provide? I'm going to go with no.
On the other hand, the existing system is mono compatible, the increased costs are slight and a mostly one time expense - you don't even have to worry about the stability of the BFO in the receiver, because you get it from the transmitter and double it.
The system in place seems more brilliant than flawed.
If you take a look at DAB in the UK, most channels are in mono!
There is no need for more than 18kHZ band width as it becomes inaudible. That is why the 19kHZ pilot was used and also as a matter of interest, the reason why Sinclair chose 22kHZ as the modulation frequency of the X10 amplifier.
There is no need for more than 18kHZ band width as it becomes inaudible. That is why the 19kHZ pilot was used and also as a matter of interest, the reason why Sinclair chose 22kHZ as the modulation frequency of the X10 amplifier.
Of course the system I described isn't compatible with mono, because we were speculating on what system might have been chosen if FM mono had not been invented first. Given the constraint of having to stay compatible with FM mono, the present stereo system was a very clever invention, despite the horrible noise penalty.
With a system with two closely spaced carrier frequencies, on the receiver side, you would just need a single LNA, single LO, single mixer and most of the IF chain could also be single. At the very end of the IF chain, you would need two narrower filters (which only need a very limited stopband suppression) to split the IF signal into two channels. The detectors and audio parts would indeed have to be doubled. That doesn't sound more complicated than the system we have now.
By the way, with DAB you have 1536 modulated carriers that are close to each other in frequency.
Last edited:
I agree with DF96. You can just add the signals of two modulators and pass them through a common PA. It's a pity that the signal is no longer constant envelope, which complicates PA design, but the same holds for AM broadcasting and analogue television (as well as DAB and DVB-T), so that wouldn't have been a showstopper.
At lunch, I happened to be looking at a paper copy of the November, 1955, issue of Radio and Television News ( 13 watt Infinite Feedback Amplifier ), but it also had an article on the then new matter of FCC approval of FM Multiplexing.
Apparently, in those days, having an FM station was a near sure fire way of turning a large fortune into a small one.
The author noted: "It is difficult to resist the conclusion that stereophonic FM in the home is an inevitable outcome of multiplex, although it is a matter of speculation how long this will take."
The article is at p. 55 if anyone is interested: https://www.americanradiohistory.com/Archive-Radio-News/50s/Radio-News-1955-11-R.pdf
Apparently, in those days, having an FM station was a near sure fire way of turning a large fortune into a small one.
The author noted: "It is difficult to resist the conclusion that stereophonic FM in the home is an inevitable outcome of multiplex, although it is a matter of speculation how long this will take."
The article is at p. 55 if anyone is interested: https://www.americanradiohistory.com/Archive-Radio-News/50s/Radio-News-1955-11-R.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.